On Apr 25, 2014, at 5:31 AM, Pete Rittwage <peter@rittwage.com> wrote: > On Fri, April 25, 2014 6:13 am, smf wrote: >>> I'm pretty sure that the last time I tested it Jiffy was slightly faster >>> than fast serial when both were used with a drive that was not limited >>> by the access/read time of the medium. >> >> I guess commodore dropped the ball again as the 1571 connected to a >> zoomfloppy can run really fast. > > It exploits a high divider/serial rate that the original 64/128 could > never hope to achieve. I think a C128 might be able to. I came up with the idea to bump the 1571’s 6502 up to 2 MHz and choose the smallest divider that the 6526A CIA supports, resulting in a 2 microsecond interval (500 Kbit/sec or 62.5 KB/sec). However, it wasn’t clear if the drive’s clock would be stable enough or if crosstalk in the IEC cable would prevent it from working reliably in practice. I show a scope trace of this here: http://www.root.org/~nate/c64/xum1541/ZoomFloppy20100918.pdf The hard work of making the drive code properly read GCR bytes and shovel them off to the CIA in time was done by Arnd Menge. He also did the reverse direction, allowing us to write out entire tracks a byte at a time. It’s pretty stable, though Pete found a couple CIAs that were marginal and didn’t work. The C128 in 2 MHz mode using its CIA might have been able to keep up, but a C64 definitely couldn’t. -Nate Message was sent through the cbm-hackers mailing listReceived on 2014-05-04 23:00:03
Archive generated by hypermail 2.2.0.